Resynchronization colorspace support failed -- premature termination of MosaikBuild when EOF is reached in one of the colorfasta mate tag files

[GoogleCodeExporter]

* What steps will reproduce the problem?

MosaikBuild -fr mate1.csfasta -fr2 mate2.csfasta \
-fq mate1.qual -fq2 mate2.qual \
-out myreads.dat -st solid -tp 1

* What is the expected output? What do you see instead?

"myreads.dat" is written but it is not completed (MosaikAligner cannot 
find any reads in this file although the file size is > 8 GBytes).
Filtering statistics is not written.
MosaikBuild stops instead with the following error message:
"ERROR: Resynchronization colorspace support failed."

My *.csfasta files do not seem to be corrupted. However, not all mate 
pairs have both mate1 (F3) and mate2 (R3) tags. This correctly triggers 
resynchronization but in my case the end of the F3 cfasta file 
(mate1.csfasta) is reached while reading F3 tags in order to synch them 
with a larger R3 tag ID found earlier, and therefore the program exits.

IMHO in such a case the program should realize that no more reads can be 
stored (the last unpaired R3 tag should be thrown away -- or treated as 
orphan), and finish the run (calculating the filtering statistics etc.) 
smoothly instead of terminating prematurely.

* What version of the product are you using? On what operating system?

Mosaik 1.0 (compiled from source -- the x64 binary does not work 
because "kernel is too old")
SLES 10 SP2 (x86_64)

* Please provide any additional information below.

Original issue reported on code.google.com by gabo...@gmail.com on 2 Nov 2009 at 8:53

[GoogleCodeExporter]

Michael,

I think I could fix the error.
The patch is in the attached context diff file.

Since I am not a C++ programmer, please check if the modification does not break
anything.

Best regards,
Gabor Toth

Original comment by gabo...@gmail.com on 2 Nov 2009 at 7:31

Attachments:

• MosaikBuild.cpp.diff

[GoogleCodeExporter]

Thanks Gabor,

I'll take the diff and integrate into the subversion tree sometime this 
weekend. 

On the other hand, colorspace conversion is a little bit shaky at the moment. 
One of
my colleagues (Wan-Ping) is working full time on the matter and a solution 
should be
ready in a week or two.

Cheers,

// Michael

Original comment by snowneb...@gmail.com on 16 Jan 2010 at 2:21

• Changed state: Accepted

[GoogleCodeExporter]

I also ran into the same problem with my mate-pair color space data. Applying 
the
suggested diff did not solve the problem. Is there an estimate when an updated 
Mosaik
version might be available?

Original comment by n.r.str...@gmail.com on 10 Mar 2010 at 11:56